Environ. Sci. Technol. 2007, 41, 785-791
Air-Water Exchange and Dry Deposition of Polybrominated Diphenyl Ethers at a Coastal Site in Izmir Bay, Turkey BANU CETIN AND MUSTAFA ODABASI* Department of Environmental Engineering, Faculty of Engineering, Dokuz Eylul University, Kaynaklar Campus, 35160 Buca, Izmir, Turkey
The air-water exchange of polybrominated diphenyl ethers (PBDEs), an emerging class of persistent organic pollutants (POPs), was investigated using paired air-water samples (n ) 15) collected in July and December, 2005 from Guzelyali Port in Izmir Bay, Turkey. Total dissolvedphase water concentrations of PBDEs (∑7PBDEs) were 212 ( 65 and 87 ( 57 pg L-1 (average ( SD) in summer and winter, respectively. BDE-209 was the most abundant congener in all samples, followed by BDE-99 and -47. Average ambient gas-phase ∑7PBDE concentrations were between 189 ( 61 (summer) and 76 ( 65 pg m-3 (winter). Net air-water exchange fluxes ranged from -0.9 ( 1.0 (BDE28) (volatilization) to 11.1 ( 5.4 (BDE-209) ng m-2 day-1 (deposition). The BDE-28 fluxes were mainly volatilization while the other congeners were deposited. Gas- and dissolved-phase concentrations were significantly correlated (r2 ) 0.33-0.55, p < 0.05, except for BDE-209, r2 ) 0.05, p > 0.05) indicating that the atmosphere controls the surface water PBDE levels in this coastal environment. Estimated particulate dry deposition fluxes ranged between 2.7 ( 1.9 (BDE-154) and 116 ( 84 ng m-2 day-1 (BDE-209) indicating that dry deposition is also a significant input to surface waters in the study area.
Introduction Polybrominated diphenyl ethers (PBDEs) are flame retardants used in several products to improve their fireproof properties (1-2). PBDEs are persistent, bioaccumulating environmental pollutants. Because of their relatively low water solubilities and vapor pressures, and their relatively large octanol-water and octanol-air partition coefficients, PBDEs are likely to partition to solids (i.e., sediment, soil, and atmospheric particles) and bioaccumulate when they are released into the environment (2). Primary emissions of persistent organic pollutants (POPs) into the atmosphere have changed over the past decades. Polychlorinated biphenyl (PCB) emissions, for example, have declined since the 1960s/1970s (3). However, as a result of their current use, PBDE concentrations are increasing in the environment (4). The dynamics of air-ocean exchange and processes within the ocean are critical to the global fate and behavior of POPs. The capacity of surface waters to store POPs is spatially and temporarily variable and is influenced by the temperature, mixing depth, and biogeochemical processes * Corresponding author phone: 90-232-412 7122; fax: 90-232412 7280; e-mail:
[email protected]. 10.1021/es061368k CCC: $37.00 Published on Web 01/03/2007
2007 American Chemical Society
(3). POPs deposited to surface waters may be further subject to incorporation into the marine food chain, degradation, and eventually deposition into the deep sea. Surface waters may therefore act as “buffers” between the atmosphere and deep sea (3). The magnitude and direction of air-water exchange has been investigated extensively because of its significant contribution to the cycling of POPs in the environment (5-9). Paired or unpaired air and water concentration measurements in conjunction with the twofilm model have been commonly used to estimate atmospheric loadings of POPs to water bodies or their loss by volatilization (5-10). However, the air-water exchange of PBDEs has not been investigated using paired samples. The objectives of this study were (1) to determine the magnitude and direction of air-water exchange fluxes of PBDEs at a coastal site in Izmir Bay, Turkey, (2) to investigate their dynamic air-water coupling, and (3) to estimate the particulate dry deposition fluxes. Paired air and water concentrations of seven PBDEs (BDE-28, -47, -99, -100, -153, -154, and 209) were measured during two sampling programs (summer and winter) in 2005. Measured concentrations and meteorological parameters were incorporated into the twofilm model to determine the magnitude and direction of airwater exchange fluxes. Particulate dry deposition PBDE fluxes were also estimated using ambient concentrations and assumed deposition velocities.
Materials and Methods Air and Water Sampling. Izmir metropolitan city, with the 2.7 million population, is the center of a highly industrialized area by the Aegean Sea shoreline of Turkey (Figure 1). Izmir is densely populated over two narrow flat basins between three series of mountains aligned perpendicular to the seashore. Concurrent surface water and air samples (n ) 15) were collected during two sampling programs on July 6-13 and December 20-26, 2005 from Guzelyali Port. Guzelyali Port is an urban site located at the south of Izmir Bay (Figure 1). Meteorological data (air temperature, wind speed and direction, and relative humidity) was taken from Guzelyali station located ∼500 meters from the sampling site while surface water temperatures were measured on-site. Average water temperatures were 12.0 ( 0.8 °C and 27.0 ( 0.8 °C for winter and summer sampling periods, respectively, while the air temperatures measured during the same periods were 5.3 ( 2.1 °C and 31.8 ( 1.5 °C. Wind speed ranged between 1.8 and 5.5 and between 2.7 and 5.2 m s-1 during the winter and summer sampling periods, respectively. Long-term observations indicate that the prevailing wind direction in the area is NW. During the sampling programs, generally northerly winds were prevailed (except the last 2 days of winter sampling) indicating that the sampled air was off the Bay water but also affected by the urban plume from the densely populated areas around the Izmir Bay (Figure 1). Water samples (∼55 L) were collected at the beginning of each air sampling (at 9.00 a.m.) manually from 30-cm depth using high-density polyethylene containers without leaving a headspace. Air samples were collected using a modified high-volume sampler (Model GPS-11, Thermo-Andersen Inc.). Particles were collected on 10.5-cm-diameter quartz filters and the gas-phase compounds were collected in a cartridge containing polyurethane foam (PUF) (6-cm diameter, 10-cm length). The average sampling volume was 96 ( 22 m3 and the sampling duration was 7 ( 1.5 h. Prior to sampling, quartz filters were baked at 450 °C overnight. Then, they were allowed to cool to room temVOL. 41, NO. 3, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
785
FIGURE 1. Map of Izmir Bay showing the sampling site. Dashed line is the border of densely populated areas. Wind rose shows the frequency (%) of prevailing wind directions during the sampling programs. perature in a desiccator. PUF cartridges were cleaned by Soxhlet extraction using acetone hexane mixture (1:1) for 24 h, were dried in an oven at 70 °C, and were stored in glass jars capped with Teflon-lined lids. After sampling, PS-1 filters and PUF cartridges were stored at -20 °C in their containers. Sample Preparation and Analysis. Water samples were filtered through a glass fiber filter (47-mm diameter) to collect particle phase in series with a resin column (∼10 g XAD-2) to collect dissolved-phase PBDEs. Four different samples were obtained in each sampling day (i.e., PUF cartridge, air filter, resin, and water filter). PUF cartridges were Soxhlet extracted for 24 h, and quartz filters, resin, and water filters were ultrasonically extracted for 60 min with a mixture of 1:1 acetone:hexane. Prior to extraction, all samples were spiked with surrogate standard (BDE-77, 3,3′,4,4′-tetrabromodiphenyl ether) to monitor the analytical recovery efficiency. The volume of extracts was reduced and transferred into hexane using a rotary evaporator and a high-purity N2 stream. After concentrated to 2 mL, sodium sulfate was added to samples to remove residual water, and samples were cleaned up and fractionated on an alumina-silicic acid column containing 3 g silicic acid (4.5% DI water) and 2 g alumina (6% DI water). The column was prewashed with 20 mL dichloromethane (DCM) followed by 20 mL petroleum ether (PE). Then, the sample in 2 mL hexane was added to the column and PBDEs were eluted with 35 mL PE. The final extracts were solvent exchanged into hexane and were concentrated to 1 mL under a stream of N2. All the samples were analyzed for PBDEs with an Agilent 6890N gas chromatograph (GC) equipped with a mass selective detector (Agilent 5973 inert MSD) working at electron capture negative chemical ionization (ECNI) mode. A capillary column (DB-5ms, 15 m, 0.25 mm, 0.1 µm) was used. The column selection and instrumental parameters were based on recent studies on the effect of the injection technique and the column system on gas chromatographic determination of PBDEs (11, 12). Pulsed-splitless injection was used to maximize the transfer of the PBDEs into the capillary column and to minimize their degradation in the injector liner. The carrier gas (helium) was used at constant flow mode (1.8 mL min-1) with a linear velocity of 70 cm s-1. The initial oven temperature was held at 90 °C for 1 min, was raised to 340 °C at 20 °C min-1, and was held for 2 min. The injector, ion source, and quadrupole temperatures were 280, 230, and 150 °C, respectively. High-purity methane was the 786
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 3, 2007
reagent gas. The MSD was run in selected ion monitoring mode. For six BDEs, the two bromine ions at m/z 79.1 and 81.1 and for BDE-209 ions at m/z 488.5 and 486.5 were monitored. Compounds were identified on the basis of their retention times and target and qualifier ions and were quantified using the internal standard calibration procedure. Quality Control. All samples were spiked with BDE-77 prior to extraction to determine the analytical recovery efficiencies. Average recovery of BDE-77 was 90 ( 13% (n ) 68) for all sample matrices. The recoveries of target compounds were also tested externally (n ) 6) and ranged between 86 ( 9% (BDE-47) and 110 ( 7% (BDE-154) (overall average ( SD, 99 ( 12%). Instrumental detection limits (IDL) were determined from linear extrapolation from the lowest standard in calibration curve using the area of a peak having a signal/noise ratio of 3. The quantifiable PBDE amounts were between 0.05 (BDE28) and 0.35 pg (BDE-209) for 1-µL injection. Blank PUF cartridges and filters were routinely placed in the field to determine if there was any contamination during sample handling and preparation. The limit of detection of the method (LOD, ng) was defined as the mean blank mass plus three standard deviations (LOD ) mean blank value + 3SD). Instrumental detection limit was used for the compounds that were not detected in blanks. LODs for six BDE congeners including BDE-28, 47, 99, 100, 153, and 154 ranged between 0.05 and 0.26 ng for water filters, between 0.05 and 0.33 ng for resins, between 0.05 and 0.34 ng for PUFs, and between 0.05 and 0.75 ng for PS-1 filters. LODs for BDE-209 were between 0.35 and 1.26 ng for water filters, resins, PUFs, and PS-1 filters. Blank amounts for water filters, resins, PUFs, and PS-1 filters were 8 ( 5%, 17 ( 10%, 13 ( 10%, and 14 ( 10% of the sample amounts, respectively. Sample quantities exceeding the LOD were quantified and blank-corrected by subtracting the mean blank amount from the sample amount. Six congeners excluding BDE-209 were also analyzed using an HP5-ms column (30 m, 0.25 mm, 0.25 µm). Since BDE209 could not be analyzed with this column, all samples were reanalyzed using a shorter DB5-ms column with a lower film thickness (15 m, 0.25 mm, 0.1 µm). Covaci et al. (13) reported the coelution of PBDEs with some other organic compounds in their study, especially coelution of hexabromobiphenyl (PBB-153) and BDE-154 and of BDE-153 and tetrabromobisphenol-A (TBBP-A). They have suggested that the column should be tested for possible coelution of target compounds. Hites (4) has suggested two different GC measurements to separate PBB-153 from BDE-154 (long column) and to measure BDE-209 (short column, preferably 10-15 m) because of its susceptibility for degradation in the GC system. In the present study, during the two different GC measurements, coelution with any other compounds and systematic differences in PBDE concentrations were not observed. The agreement between the two sets of analysis was very good (statistically the same at 95% confidence level, two-tailed t-test) indicating the reliability of GC techniques used in this study. Possible artifacts associated with the high-volume air sampling are the adsorption of gas-phase chemical to the filter and the breakthrough of the gas-phase compounds through the PUF. These artifacts were recently investigated for BDEs using sampling volumes ranging between 300 and 600 m3 (outdoor) and between 100 and 200 m3 (indoor) (14). It was shown that the sampling artifacts were insignificant for the measured total BDE concentrations (gas + particle phases, sum of nine congeners) ranging between 39 (outdoor) and 2088 pg m-3 (indoor) (14). In the present study, since the sampling volume (96 m3) was relatively low, and the maximum total BDE concentration was approximately an order of magnitude lower than the one reported by Shoeib et al. (14), it was assumed that the PUF effectively captured
TABLE 1. Concentrations of PBDEs in Water and Air (Average ( SD) water concentrations (pg L-1) dissolved phase
air concentrations (pg m-3)
particle phase
gas phase
particle phase
congenera summer (n ) 8) winter (n ) 7) summer (n ) 8) winter (n ) 7) summer (n ) 8) winter (n ) 7) summer (n ) 8) winter (n ) 7) BDE-28 BDE-47 BDE-100 BDE-99 BDE-154 BDE-153 BDE-209 ∑7BDEs
2.7 ( 1.4 40 ( 9 11 ( 3 57 ( 15 5.4 ( 1.3 8.2 ( 2.1 89 ( 52 212 ( 65
1.4 ( 0.9 9.5 ( 3.0 3.4 ( 1.2 14 ( 5 1.6 ( 0.9 2.5 ( 1.1 54 ( 46 87 ( 57
3.1 ( 1.0 34 ( 7 11 ( 2 53 ( 10 5.7 ( 0.8 8.3 ( 1.1 157 ( 47 271 ( 60
1.5 ( 0.9 11 ( 5 3.8 ( 1.5 16 ( 7 1.8 ( 0.6 2.7 ( 1.5 443 ( 329 479 ( 340
3.4 ( 1.7 49 ( 20 11 ( 3 63 ( 22 5.2 ( 1.7 7.3 ( 1.7 49 ( 15 189 ( 61
1.0 ( 0.4 6.9 ( 5.9 2.9 ( 2.0 11 ( 7 1.1 ( 0.6 1.2 ( 0.5 52 ( 52 76 ( 65
1.1 ( 0.7 12 ( 7 4.5 ( 2.0 23 ( 11 2.7 ( 1.1 3.4 ( 1.5 36 ( 19 83 ( 37
0.6 ( 0.6 6.7 ( 4.8 2.7 ( 1.7 9.7 ( 5.9 1.0 ( 0.8 1.0 ( 0.9 46 ( 34 68 ( 48
a 2,4,4′-Tribromodiphenyl ether (BDE 28), 2,2′,4,4′-tetrabromodiphenyl ether (BDE 47), 2,2′,4,4′,5-pentabromodiphenyl ether (BDE 99), 2,2′,4,4′,6pentabromodiphenyl ether (BDE 100), 2,2′,4,4′5,5′-hexabromodiphenyl ether (BDE 153), 2,2′,4,4′5,6′-hexabromodiphenyl ether (BDE 154), and decabromodiphenyl ether (BDE 209) (PBDE congeners are numbered according to IUPAC system established for PCBs).
the gas-phase fraction. The possibility of chemical breakthrough from XAD-2 adsorbent trap was also checked by connecting a second trap in series. No breakthrough was observed. Air-Water Exchange Modeling. According to the twofilm model, mass transfer is limited by the rate of molecular diffusion through thin films of air and water on either side of the surface (15). The gas flux across a water surface is a function of Henry’s law constant, the concentration gradient, and the overall mass transfer coefficient (10, 15). The net flux (Fg, ng m-2 day-1) is driven by the fugacity difference between air and surface water:
Fg ) Kg (Cg - CwH/RT)
(1)
where Cw and Cg are the water and air concentrations (ng m-3), H is the Henry’s law constant (Pa m-3 mol-1), R is the universal gas constant (8.314 Pa m3 mol-1 K-1), and T is temperature at the air-water interface (K). The gas-phase overall mass transfer coefficient (Kg) is related to individual mass transfer coefficients for the liquid and gas films, kw and ka, as follows:
1/Kg ) (1/ka) + (H/RTkw)
(2)
Mass transfer coefficients of water vapor, oxygen (O2), and carbon dioxide (CO2) have been related to wind speed by many researchers The following equations can be used to estimate ka and kw for organic compounds (15, 16):
ka(compound) (cm s-1) ) (0.2U10 + 0.3) [Da(compound)/Da(H2O)]0.67 (3) kw(compound) (cm s-1) ) [(0.24U102 + 0.061U10)/3600] [Dw(compound)/Dw(CO2)]0.5 (4) where Da and Dw (cm2 s-1) are the diffusivities in air and water, respectively, and U10 is the wind speed 10 m above the water surface (m s-1).
Results and Discussion Water Concentrations. Total dissolved-phase PBDE concentrations (∑7PBDEs) were 212 ( 65 and 87 ( 57 pg L-1 (average ( SD) for summer and winter seasons, respectively, and they were 271 ( 60 and 479 ( 340 pg L-1 for particlephase concentrations (Table 1). For six BDE congeners, including BDE-28, 47, 99, 100, 153, and 154, dissolved- and particle-phase concentrations were similar during the sampling periods. However, for BDE-209, particle-phase concentrations were ∼8 times higher than dissolved-phase concentrations in winter. BDE-209 was the dominating
congener in both phases and periods, followed by BDE-99 and 47. In summer, BDE-209 contributed 42 and 52% to total dissolved- and particle-phase concentrations, respectively. In winter, its contribution to total dissolved- and particlephase concentrations was 63 and 92%, respectively. Because of the abundance of BDE-209, the data set was divided into two subsets as BDE-209 and ∑6BDEs (Figure 2). Particlephase water BDE-209 concentrations were significantly higher especially in winter periods (Figure 2). The distribution of PBDEs between dissolved and particulate fractions were similar in winter and summer periods, except BDE-209. Particulate fractions of six BDE congeners were 50 and 53% for summer and winter periods, respectively (Figure 2). However, particulate fractions for BDE-209 were 64% and 89% for summer and winter periods, respectively, indicating its predominant partition into particulate matter. Recently, Oros et al. (17) reported that the ∑PBDEs were mostly partitioned into the suspended particulate matter fraction with the relative abundances between 78 and 93% of total PBDE concentrations. In the present study, only the partition of BDE-209 was similar to this observation. There is limited information on PBDEs in environmental waters. The sum of dissolved- and particle-phase individual PBDE concentrations ranged between 0.2 and 191 pg L-1 and BDE-47, 99, and 209 were the dominating congeners in the San Francisco Estuary (17). BDE-28, 47, 100, and 183 were the dominating congeners while BDE-209 was detected only in trace amounts at five sampling locations in Hong Kong (18). However, in most locations, PBDE concentrations were low or below detection limit. The mean dissolved- and particle-phase seawater concentrations for eight congeners were 71 and 28 pg L-1, respectively, and they were 149 and 38 pg L-1 for sea-surface microlayer (18). Ambient Air Concentrations. Gas-phase ∑7PBDE concentrations were 189 ( 61 and 76 ( 65 pg m-3 (average ( SD), for summer and winter periods, respectively (Table 1). Gas-phase BDE-209 did not show a seasonal trend as the other six BDEs (Figure 2). Multiple linear regression analysis (19) indicated that concentrations increased with temperature and the correlation was statistically significant for six BDEs excluding BDE-209 (p < 0.01, r2 ) 0.69-0.87). However, BDE concentrations were not correlated to wind speed and direction. Particle-phase PBDE concentrations were 83 ( 37 and 68 ( 48 pg m-3 for summer and winter periods, respectively. BDE-209, 99, and 47 were the dominating congeners in all samples, contributing up to 85-92% to total PBDE concentrations (Table 1). In summer, gas-phase PBDEs were dominated by BDE-99 (34%), followed by BDE-47 and 209 while in particle-phase, 42% of PBDEs consisted of BDE209. However, in winter, the contribution of BDE-209 to total concentrations was 59 and 67%, for gas and particle phases, respectively (Table 1). VOL. 41, NO. 3, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
787
FIGURE 2. Water (particle and dissolved) and ambient air (particle- and gas-phase) concentrations of BDE-209 and ∑6BDE for summer and winter 2005 sampling periods. *, not analyzed. The concentrations of PBDEs measured in the present study were within the ranges of previously reported values. Gas-phase concentrations of ∑7PBDEs measured recently at another urban site in Izmir, Turkey were 13 ( 5 and 4.6 ( 2.1 pg m-3, for summer and winter periods, respectively, and for particle phase they were 25 ( 12 and 36 ( 17 pg m-3 (20). Similar to the present study, BDE-209 was the most abundant congener (20). A recent study conducted using passive samplers at a continental scale reported that ∑8PBDE concentrations (excluding BDE-209) ranged between 0.5 and 250 pg m-3 over Europe (21). In southern Ontario, BDE-209, 47, and 99 concentrations were up to 105, 10, and 16 pg m-3, respectively (22, 23). Hoh and Hites (24) have reported that the average atmospheric ∑PBDE concentration at the Chicago site was 100 ( 35 pg m-3, lower than the one measured in the present study. Hoh and Hites (24) found that BDE-47 was the dominating congener and the average fraction of BDE-209 was only 6-31%, lower than those observed in the present study. Trace BDE-209 levels were measured only at the Chicago site while BDE-47 and 99 accounted for 50-65% and 35-40% of the total PBDE concentrations, respectively, at urban, rural, and remote sites near the Great Lakes between 1997 and 1999 (1). Chen et al. (25) have reported that the average ∑11PBDE concentration (particle + gas) in an urban site of China was 353 pg m-3. They have found that BDE-209 was the dominating congener with 40-99% abundance. Chen et al. (25) have suggested that the absence of BDE-209 in some previous studies may be due to an analytical bias. Gas-particle partitioning models estimate that higher brominated congeners like BDE-209 are expected to be particle bound >99% (20). However, Agrell et al. (26) have found that BDE-209 was mainly in gas phase (>90%) at an urban site while it was 100% in gas phase at a rural site. In southern Ontario, BDE-209 was found only in particle phase (22). It was suggested that the reason for not detecting BDE209 in gas phase may be its sorption to the glass fiber filter during sampling (22). The average proportion of BDE-209 in particle phase was 70 ( 22% at four different sites in Izmir, Turkey (20). A slightly lower value (50 ( 24%) was observed in the present study that lies between the extreme values (0-100% particle phase) reported by Agrell et al. (26) and Gouin et al. (22). 788
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 3, 2007
Air-Water Exchange. Concurrent air and water concentrations are ideally used to assess the state of equilibrium for individual POPs between the air-water interfaces. The water-air fugacity ratio (fW/fA ) H′Cw/Cg) > 1.0 indicates net volatilization of compounds from water, values
